Computing confidence level in spectrum (after fft).

Hello,

I would like to compute confidence levels after I did a fft. What are the approaches? I tried:

1) to compute the background of the spectrum with a quadratic fit (in log-Power-spec). Afterwards I want to add the confidence interval (calculated with the variance). - Here the fit looks quite bad.

2) to use pwelch. With [pxx,f_C,pxxc] = pwelch(x,[],[],fs,'ConfidenceLevel',0.95) I get some errors:

   Error using zeros
   Size inputs must be integers.

   Error in welch (line 81)
   Sxx = zeros(nFreqs,size(x,2),class(x)); %#ok<*ZEROLIKE>

   Error in pwelch (line 158)
   [varargout{1:nargout}] = welch(x,esttype,varargin{:});

   Error in preproc_ngk (line 701)
   [pxx,f_C,pxxc] = pwelch(x,hamming(L),noverlap,fs,'ConfidenceLevel',0.95);

Maybe you know a different way to solve the problem? Or maybe there is a mistake with pwelch.

Thanks, Jonah

[Jie Zhang]

I have the same question. Could you please tell me how you solved it? Thanks a lot.

[Jonah Geils]

Hi, I didn't solved it at all -_-